Macrophage depletion induces edema through release of matrix-degrading proteases and proteoglycan deposition.

Colony-stimulating factor 1 receptor (CSF1R) blockade abates tumor-associated macrophage (TAM) infiltrates and provides marked clinical benefits in diffuse-type tenosynovial giant cell tumors. However, facial edema is a common adverse event associated with TAM elimination in patients. In this study, we examined molecular and cellular events associated with edema formation in mice and human patients with cancer treated with a CSF1R blocking antibody. Extended antibody treatment of mice caused marked body weight gain, an indicator of enhanced body fluid retention. This was associated with an increase of extracellular matrix-remodeling metalloproteinases (MMPs), namely MMP2 and MMP3, and enhanced deposition of hyaluronan (HA) and proteoglycans, leading to skin thickening. Discontinuation of anti-CSF1R treatment or blockade of MMP activity restored unaltered body weight and normal skin morphology in the mice. In patients, edema developed at doses well below the established optimal biological dose for emactuzumab, a CSF1R dimerization inhibitor. Patients who developed edema in response to emactuzumab had elevated HA in peripheral blood. Our findings indicate that an early increase of peripheral HA can serve as a pharmacodynamic marker for edema development and suggest potential interventions based on MMP inhibition for relieving periorbital edema in patients treated with CSF1R inhibitors.

The Tumor Milieu Promotes Functional Human Tumor-Resident Plasmacytoid Dendritic Cells in Humanized Mouse Models.

Particular interest to harness the innate immune system for cancer immunotherapy is fueled by limitations of immune checkpoint blockade. Plasmacytoid dendritic cells (pDC) are detected in a variety of solid tumors and correlate with poor clinical outcome. Release of type I interferons in response to toll-like-receptor (TLR)7 and TLR9 activation is the pDC hallmark. Mouse and human pDC differ substantially in their biology concerning surface marker expression and cytokine production. Here, we employed humanized mouse models (HIS) to study pDC function. We performed a comprehensive characterization of transgenic, myeloid-enhanced mouse strains (NOG-EXL and NSG-SGM3) expressing human interleukin-3 (hIL-3) and granulocyte-macrophage colony stimulating factor (GM-CSF) using identical humanization protocols. Only in HIS-NOG-EXL mice sufficient pDC infiltration was detectable. Therefore, we selected this strain for subsequent tumor studies. We analyzed pDC frequency in peripheral blood and tumors by comparing HIS-NOG-EXL with HIS-NOG mice bearing three different ovarian and breast tumors. Despite the substantially increased pDC numbers in peripheral blood of HIS-NOG-EXL mice, we detected TLR7/8 agonist responsive and thus functional pDCs only in certain tumor models independent of the mouse strain employed. However, HIS-NOG-EXL mice showed in general a superior humanization phenotype characterized by reconstitution of different myeloid subsets, NK cells and B cells producing physiologic IgG levels. Hence, we provide first evidence that the tumor milieu but not genetically introduced cytokines defines intratumoral (i.t.) frequencies of the rare pDC subset. This study provides model systems to investigate in vivo pro- and anti-tumoral human pDC functions.